Response of semi-permanent prairie wetland to climate change: a spatial simulation model

The objective of this research was to assess the potential effects of global warming on the hydrology and vegetation in semi-permanent wetlands located in the glaciated prairie region of North Dakota. As a means to that objective, a spatially-defined simulation model of the vegetation dynamics in these wetlands was constructed.

A hydrologic component of the model estimated water levels based on precipitation, runoff, potential evaporation and transpiration. Amount and distribution of emergent cover and open water were modeled using a geographical information system. Vegetation response to changes in water level was based on seed bank composition, seedling recruitment, establishment and plant survivorship. Simulation results were compared to actual distributions from aerial photographs (1979-89).

Results showed that the model was relatively good at calculating changes in water level for average years. Late-summer water levels were overestimated during dry years due to limitations in the Thornthwaite method of calculating potential evapotranspiration.

In general, changes in the ratio of emergent cover to open water were accurately simulated. Tests of the model elucidated two areas that needed improvement. First, seedlings germinated too quickly on exposed mudflats in the model when drawdown occurred late in the season. The actual wetland had a thick mat of dried, submergent vegetation on top of the mudflats which impeded germination, which the model did not consider. Second, model conversions between open water and deep marsh vegetation were not always timed correctly. If water depth crossed a threshold value for a given period of time a cell would change its type. In reality, tolerance of emergents to deep water is more complex. A probability function with respect to time and water depth rather than a threshold value would better represent this relationship.

The model was used to assess the potential effects of global warming on the cover cycle in one wetland. An 11-year simulation was run using a normal versus greenhouse climate. Although water level fluctuations still occurred, peak values were significantly lower in the warming scenario and the wetland dried in most years. Simulations also revealed a significant change in the vegetation, from a nearly balanced cover ratio to a completely closed basin with no open water areas.